MancJiester Memoirs, Vol. xlii. ( 1 898), No. 6. 45 



{a) Balancing the brake. 



This was an extremely difficult operation to perform, 

 owing to the large amount of friction which occurred 

 between the brake and the engine shaft. A number of 

 experiments were made which shewed that the want of 

 balance could not exceed ^ ft.-lb. This want of balance 

 would make an error in the work calculated for each 

 individual trial, but on the difference of work done in any 

 two trials the error was completely eliminated if each had 

 the same number of revolutions. Consequently, the speed 

 of the engine was regulated to the end that the number of 

 revolutions should be the same in any two trials which 

 were afterwards to be compared. 



Even with a difference of 300 revolutions, which was 

 about the maximum difference, between the two trials, 

 the error of yi ft.-lb. in balance would make an error of 

 less than 00015 per cent in the calculated difference of 

 work. Further, since this error was a casual one, it was 

 probable that it would cut out on the mean final result, 

 and being so extremely small on any individual determina- 

 tion it has been neglected. 



ib) Length of the lever. 



This was determined by very careful observation to 

 be 4ft. -|-o-02in. when the engine was running under all the 

 conditions which obtained in the trials. As in the pre- 

 liminary calculations this length was given its nominal 

 value of 4ft., the value of K obtained required a correction 

 of -J-0042 per cent. 



(r) The zveights used in loading the brake. 



These were cast iron plates weighing 25lbs. each. 

 Their weights were determined to rJ-Tj-lb., and in the later 

 trials they were so arranged that no error could appear 

 in the final result on account of the slight differences 



